Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Index Backfill : Bubble up failed master -> TS rpcs. #3627

Closed
amitanandaiyer opened this issue Feb 12, 2020 · 0 comments
Closed

Index Backfill : Bubble up failed master -> TS rpcs. #3627

amitanandaiyer opened this issue Feb 12, 2020 · 0 comments
Assignees

Comments

@amitanandaiyer
Copy link
Contributor

Part of/Related to #448

@amitanandaiyer amitanandaiyer self-assigned this Feb 12, 2020
amitanandaiyer added a commit that referenced this issue Feb 14, 2020
Summary:
1) Handle failed master -> TS rpcs. These failures
should bubble up and abort the index creation process.

Fix  RetryingTSRpcTask to not fast-fail and abort when a TS is not found. Instead, it will
backoff and wait until deadline to see if a TS for that tablet registers with the master.

update fully_applied_* fields once all tablets have moved to the
backfill phase. This is to minimise schema version mismatch errors
 while the backfill is going on.

Crisper description for monitored tasks. Useful to monitor
the progress at http://<master>:7000/tasks

use index_ids in description to make the description less verbose.

Test Plan:
ybd --cxx-test cassandra_cpp_driver-test --gtest_filter CppCassandraDriver*.*TestCreateIndexSlowTServer*

ybd --cxx-test integration-tests_master_failover-itest --gtest_filter MasterFailoverTestIndexCreation.TestPauseAfterCreateIndexIssued

Reviewers: mihnea, rahuldesirazu, hector

Reviewed By: hector

Subscribers: ybase, bogdan

Differential Revision: https://phabricator.dev.yugabyte.com/D7943
amitanandaiyer added a commit that referenced this issue Feb 20, 2020
…uent tables.

Summary:
The very first table among colocated tables should not issue AddTableToTablet.

The issue was masked by the fact that issuing such an RPC before the CreateTablet RPC succeeded used to fail the RPC right away. The RPC was fixed to retry before giving up in D7943 (#3627). Subsequently, we've seen tservers running into DFATAL in the following tests.

./yb_build.sh --cxx-test pg_libpq-test --gtest_filter PgLibPqTest.TableColocation
./yb_build.sh --java-test org.yb.pgsql.TestPgRegressBetaFeatures

Test Plan:
./yb_build.sh --cxx-test pg_libpq-test --gtest_filter PgLibPqTest.TableColocation
./yb_build.sh --java-test org.yb.pgsql.TestPgRegressBetaFeatures

Reviewers: neha, mihnea, hector, bogdan, jason

Reviewed By: jason

Subscribers: ybase, yql

Differential Revision: https://phabricator.dev.yugabyte.com/D7987
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

1 participant